An Artificial Neural Network Model for Flexoelectric Actuation and Control of Beams

Abstract

The converse flexoelectric effect has been applied to precision actuation and vibration control of flexible structures. High stress concentration caused by a single flexoelectric actuator can be alleviated by placing multiple actuators on the structure. In the presented work, a neural network model was established to optimize the positions of multiple flexoelectric actuators on a cantilever beam. It was proved that the neural network can recognize the relationship between actuator position and tip displacement and forecast the tip displacement of the beam accurately with reduced computational effort and higher effectiveness. By using the neural network, the displacement data generated by all possible combinations of actuator positions were predicted, and the optimal positions of multiple flexoelectric actuators can be obtained in term of maximum transverse tip displacement. The effect of actuator size with various actuator numbers was discussed. The results showed that applied voltage can be reduced by increasing the number of flexoelectric actuators when placed on optimal positions.